摘要
为了准确计算颅内额叶损伤的出血面积,提出了将区域生长、FCM模糊聚类用于额叶出血CT图的分割,提取出CT图中出血的目标区域,实现出血面积的精确计算。用该方法分割出的出血区域面积与CT图中实际的出血面积相对误差在5%左右。同时将该图像分割方法与区域生长、阈值分割为一体的分割方法进行比较,发现当CT图中血块区域与周围的脑组织灰度值差异较小时,区域生长、FCM模糊聚类为一体的图像分割方法的分割结果较为精确。
In order to accurately calculate the area of bleeding injury intracranial frontal lobe, the method of region growing and FCM fuzzy clustering for the segmentation of Intracranial hemorrhage CT image was raised, target area of bleeding in the CT images was extracted accurately and calculation of the area of bleeding was achieved accurately. The relative error between the bleeding area used by this method and the actual bleeding area is about 5%; At the same time, comparing the method of region growing and FCM fuzzy clustering with the method of regional growth and threshold segmentation, it is found that when the difference of gray value between the bleeding area and the surrounding brain matte is small, the segmentation result of the method of region growing and FCM fuzzy clustering is more accurate.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2014年第2期231-235,共5页
Journal of System Simulation
基金
国家自然科学基金(61273069)
福建省产学研重大专项资助(2011H6019)
关键词
额叶损伤
出血面积
区域生长
FCM模糊聚类
图像分割
frontal lobe lesion
hemorrhage area
region growing
FCM fuzzy clustering
image segmentation
